Effect of Pervasive
Encryption
Dell EMC
176 South St
Hopkinton
MA
USA
+1
Kathleen.Moriarty@dell.com
AT&T Labs
200 Laurel Avenue South
Middletown,
NJ
07748
USA
+1 732 420 1571
+1 732 368 1192
acmorton@att.com
http://home.comcast.net/~acmacm/
Increased use of encryption impacts operations for security and
network management causing a shift in how these functions are performed.
In some cases, new methods to both monitor and protect data will evolve.
In other cases, the ability to monitor and troubleshoot could be
eliminated. This draft includes a collection of current security and
network management functions that may be impacted by the shift to
increased use of encryption. This draft does not attempt to solve these
problems, but rather document the current state to assist in the
development of alternate options to achieve the intended purpose of the
documented practices.
In response to pervasive monitoring revelations and the IETF
consensus that Pervasive Monitoring is an Attack , efforts are underway to increase encryption of
Internet traffic. Session encryption helps to prevent both passive and
active attacks on transport protocols; more on pervasive monitoring can
be found in the Confidentiality in the Face of Pervasive Surveillance: A
Threat Model and Problem Statement . The
Internet Architecture Board (IAB) released a statement advocating for
increased use of encryption in November 2014. Views on acceptable
encryption have also shifted and are documented in "Opportunistic
Security" (OS) , where cleartext sessions should
be upgraded to unauthenticated session encryption, rather than no
encryption. OS encourages upgrading from cleartext, but cannot require
or guarantee such upgrades. Once OS is used, it allows for an evolution
to authenticated encryption. These efforts are necessary to improve end
user's expectation of privacy, making pervasive monitoring cost
prohibitive. Active attacks are still possible on sessions where
unauthenticated sessions are in use. The push for ubiquitous encryption
via OS is specific to improving privacy for everyday users of the
Internet.
Although there is a push for OS, there is also work being done to
improve implementation development and configuration flaws of TLS and
DTLS sessions to prevent active attacks used to monitor or intercept
session data. The (UTA) working group is in process of publishing
documentation to improve the security of TLS and DTLS sessions. They
have documented the known attack vectors in and
have documented Best Practices for TLS and DTLS in and have other documents in the queue.
Estimates for session encryption from spring 2015 approximate that
about 30% of web sites have session encryption enabled, according to the
Electronic Frontier Foundation . The Mozilla
maintains statistics on TLS usage and as of March 2017, 54% of HTTP base
page loads are encrypted. The statistic from Mozilla varies when filters
are applied for platform and browser versions. Enterprise networks such
as EMC, now Dell EMC, observed that about 78% of outbound employee
traffic was encrypted in June 2014. Although the actual number of sites
may only be around 30%, they include some of the most visited sites on
the Internet for corporate users.
In addition to encrypted web site access (HTTP over TLS), there are
other well-deployed application level transport encryption efforts such
as mail transfer agent (MTA)-to-MTA session encryption transport for
email (SMTP over TLS) and gateway-to-gateway for instant messaging (XMPP
over TLS). Although this does provide protection from transport layer
attacks, the servers could be a point of vulnerability if user-to-user
encryption is not provided for these messaging protocols. User-to-user
content encryption schemes, such as S/MIME and PGP for email and
encryption (e.g. Off-the-Record (OTR)) for Extensible Messaging and
Presence Protocol (XMPP) are used by those interested to protect their
data as it crosses intermediary servers, preventing the vulnerability
described by providing an end-to-end solution. User-to-user schemes are
under review and additional options will emerge to ease the
configuration requirements, making this type of option more accessible
to non-technical users interested in protecting their privacy.
Increased use of encryption (either opportunistic or authenticated)
will impact operations for security and network management, causing a
shift in how these functions are performed. In some cases new methods to
monitor and protect data will evolve, for other cases the capability may
be eliminated. This draft includes a collection of current security and
network management functions that may be impacted by this shift to
increased use of encryption. This draft does not attempt to solve these
problems, but rather document the current state to assist in the
development of alternate options to achieve the intended purpose of the
documented practices.
In this document we consider several different forms of service
providers, so we distinguish between them with adjectives. For example,
network service providers (or network operators) provide IP-packet
transport primarily, though they may bundle other services with packet
transport. Alternatively, application service providers primarily offer
systems that participate as an end-point in communications with the
application user, and hosting service providers lease computing,
storage, and communications systems in datacenters. In practice, many
companies perform two or more service provider roles, but may be
historically associated with one.
Network Service Providers (SP) are responding to encryption on the
Internet, some helping to increase the use of encryption and others
preventing its use. Network SPs for this definition include the backbone
Internet Service providers as well as those providing infrastructure at
scale for core Internet use (hosted infrastructure and services such as
email).
Following the Snowden revelations, application service providers
responded by encrypting traffic between their data centers to prevent
passive monitoring from taking place unbeknownst to them (Yahoo, Google,
etc.). Large mail service providers also began to encrypt session
transport to hosted mail services. This had an immediate impact to help
protect the privacy of users data, but created a problem for some
network operators. They could no longer gain access to session streams
resulting in actions by several to regain their operational practices
that previously depended on cleartext data sessions.
The EFF reported several network service
providers taking steps to prevent the use of SMTP over TLS by breaking
STARTTLS (section 3.2 of ), essentially
preventing the negotiation process resulting in fallback to the use of
clear text. Some methods used by service providers are impacted by the
use of encryption where middle boxes were in use to perform functions
that range from load balancing techniques to monitoring for attacks or
enabling "lawful intercept", such that described in and in the US.
Network service providers use various monitoring techniques for
security and operational purposes. The following subsections detail the
purpose of each type of monitoring and what protocol fields are used to
accomplish the task. The loss of access to these fields may prompt
undesirable security practices in order to gain access to the fields in
unencrypted data flows. Ideally, new methods could be developed to
accomplish the same goal without the ability to see session data.
A standalone load balancer is something one can take off the shelf,
place in front of a pool of servers, and with an appropriate
configuration, it will load balance the traffic. This is a typical
setup that one thinks of when they think of load balancer middleboxes.
Standalone load balancers can only rely on the plainly-observable
information in the packets they are forwarding and can only rely on
the industry-accepted standards in interpreting the plainly-observable
information. Typically, this is a 5-tuple of the connection.
An integrated load balancer is developed to be an integral part of
the service provided by the server pool behind that load balancer.
These load balancers can communicate state with their pool of servers
to better route flows to the appropriate servers. They can rely on
non-standard system-specific information and operational knowledge
shared between the load balancer and its servers.
Both standalone and integrated load balancers can be deployed in
pools for redundancy and load sharing. For high availability, it is
important that when packets belonging to a flow start to arrive at a
different load balancer in the load balancer pool, the packets
continue to be forwarded to the original server in the server pool.
The importance of this requirement increases as the chances of such
load balancer change event increases.
With the proliferation of mobile connected devices, there is an
acute need for connection-oriented protocols that maintain connections
after a network migration by an endpoint. QUIC is an example of such
protocol in development by IETF QUIC WG right now. This connection
persistence provides an additional challenge for multi-homed
anycast-based services typically employed by large content owners and
Content Distribution Networks (CDNs). The challenge is that a
migration to a different network in the middle of the connection
greatly increases the chances of the packets routed to a different
anycast pop due to the new network?s different connectivity and
Internet peering arrangements. The load balancer in the new pop,
potentially thousands of miles away, will not have information about
the new flow and would not be able to route it back to the original
pop.
To help with the endpoint network migration challenges, anycast
service operations are likely to employ integrated load balancers
that, in cooperation with their pool servers, are able to ensure that
client-to-server packets contain some additional identification in
plainly-observable parts of the packets (in addition to the 5-tuple).
New connection-migration-tolerant protocols, such as QUIC, are
deliberately designed to allow such extra information available in
plain text (QUIC?s server-generated flow IDs). (That said, care must
be exercised to make sure that the information encoded by the
endpoints is not sufficient to identify unique flows and facilitate
Persistent Surveillance attack vector.)
Some integrated load balancers utilize the ability to have
additional plainly observable information even for today?s protocols
that are not network migration tolerant. This additional information
bestows advantages in additional availability and scalability to such
load balancers. For example, BGP reconvergence can cause a flow to
switch anycast pops even without a network change by any endpoint.
Additionally, a system that is able to encode the identity of the pool
server in plain text information available in each incoming packet is
able to provide stateless load balancing. This ability confers great
reliability and scalability advantages even if the flow remains in a
single pop. Indeed, a stateless load balancing system is not required
to keep state of each flow. Even more importantly, it is not required
to continuously sync such state among the pool of load balancers.
Current protocols, such as TCP, allow the development of stateless
integrated load balancers by availing such load balancers of
additional plain text information in client-to-server packets. (In
case of TCP, such information can be encoded by having
server-generated sequence numbers, mss values, lengths of the packet
sent, etc.)
In future Network Function Virtualization (NFV) architectures, load
balancing functions are likely to be more prevalent (deployed at
locations throughout operators' networks), so they would be handling
traffic using encrypted tunnels whenever it is present.
Internet traffic surveys are useful in many well-intentioned
pursuits, such as CAIDA data and SP network
design and optimization. Tracking the trends in Internet traffic
growth, from earlier peer-to-peer communication to the extensive
adoption of unicast video streaming applications, has required a view
of traffic composition and reports with acceptable accuracy. As
application designers and network operators both continue to seek
optimizations, the role of traffic survey (e.g. passive monitoring)
retains its importance.
Passive monitoring makes inferences about observed traffic using
the maximal information available, and is subject to inaccuracies
stemming from incomplete sampling (of packets in a stream) or loss due
to monitoring system overload. When encryption conceals more layers in
each packet, reliance on pattern inferences and other heuristics
grows, and accuracy suffers. For example, the traffic patterns between
server and browser are dependent on browser supplier and version, even
when the sessions use the same server application (e.g., web e-mail
access). It remains to be seen whether more complex inferences can be
mastered to produce the same monitoring accuracy.
Fingerprinting is used in traffic analysis and monitoring to
identify traffic streams that match certain patterns. This technique
may be used with clear text or encrypted sessions. Some Distributed
Denial of Service (DDoS) prevention techniques at the Network SP
level rely on the ability to fingerprint traffic in order to
mitigate the effect of this type of attack. Thus, fingerprinting may
be an aspect of an attack or part of attack countermeasures.
A common, early trigger for DDoS mitigation includes observing
uncharacteristic traffic volumes or sources; congestion; or
degradation of a given network or service. One approach to mitigate
such an attack involves distinguishing attacker traffic from
legitimate user traffic. The ability to examine layers and payloads
above transport provides a new range of filtering opportunities at
each layer in the clear. If fewer layers are in the clear, this
means that there are reduced filtering opportunities available to
mitigate attacks. However, fingerprinting is still possible.
Passive monitoring of network traffic can lead to invasion of
privacy by external actors of the endpoints of the monitored
traffic. Encryption of traffic end-to-end is one method to obfuscate
some of the potentially identifying information, it is not a
panacea, however. Many DoS mitigation systems perform this manner of
passive monitoring.
For example, browser fingerprints are comprised of many
characteristics, including User Agent, HTTP Accept headers, browser
plug-in details, screen size and color details, system fonts and
time zone. A monitoring system could easily identify a specific
browser, and by correlating other information, identify a specific
user.
Two applications of DPI are covered below, where DPI means
inspection deeper than the 5-tuple for the purpose of this document.
These applications include caching and differential treatment.
The features and efficiency of some Internet services can be
augmented through analysis of user flows and the applications they
provide. For example, network caching of popular content at a
location close to the requesting user can improve delivery
efficiency (both in terms of lower request response times and
reduced use of International Internet links when content is
remotely located), and authorized parties use DPI in combination
with content distribution networks to determine if they can
intervene effectively. Web proxies are widely used , and caching is supported by the recent update
of "Hypertext Transfer Protocol (HTTP/1.1): Caching" in . Encryption of packet contents at a given
protocol layer usually makes DPI processing of that layer and
higher layers impossible. It should be noted that some content
providers prevent caching intentionally to control content
delivery through the use of encrypted end-to-end sessions. The
business risk is a motivation outside of privacy and pervassive
monitoring that are driving end-to-end encryption for these
content providers.
Data transfer capacity resources in cellular radio networks
tend to be more constrained than in fixed networks. This is a
result of variance in radio signal strength as a user moves around
a cell, the rapid ingress and egress of connections as users
hand-off between adjacent cells, and temporary congestion at a
cell. Mobile networks alleviate this by queuing traffic according
to its required bandwidth and acceptable latency: for example, a
user is unlikely to notice a 20ms delay when receiving a simple
Web page or email, or an instant message response, but will very
likely notice a re-buffering pause in a video playback or a VoIP
call de-jitter buffer. Ideally, the scheduler manages the queue so
that each user has an acceptable experience as conditions vary,
but knowledge of the traffic type has been used to make bearer
assignments and set scheduler priority. Application and transport
layer encryption make the traffic type estimation more complex and
less accurate, and therefore it may not be effectual anymore to
use this information as input for queue management. These effects
and potential alternative solutions have been discussed at the
accord BoF at IETF95.
DPI allows identification of applications based on payload
signatures, in contrast to trusting well-known port numbers.
Operators plan network infrastructure based on demographic shifts
in application usage. Past shifts have included the growth of
peer-to-peer file sharing during all hours of the day and more
recently growth in streaming video at prime time, both of which
have impacted network design.
When called upon to diagnose customer complaints, the starting
point may be a particular application that isn't working. Being
able to identify that application's traffic using DPI is
important; IP address filtering is not useful for applications
using CDNs or cloud providers. After identifying the traffic, an
operator may analyze the traffic characteristics and routing of
the traffic.
In contrast to DPI, various applications exist to provide data
compression in order to conserve the life of the user's mobile data
plan and optimize delivery over the mobile link. The compression proxy
access can be built into a specific user level application, such as a
browser, or it can be available to all applications using a system
level application. The primary method is for the mobile application to
connect to a centralized server as a proxy, with the data channel
between the client application and the server using compression to
minimize bandwidth utilization. The effectiveness of such systems
depends on the server having access to unencrypted data flows. As the
percentage of connections using encryption increases, these data
compression services will be rendered less effective, or worse, they
will adopt undesirable security practices in order to gain access to
the unencrypted data flows.
There are numerous motivations for service proividers to block
content. See RFC7754 for a survey of
internet filtering techniques and motivations, not specific to
content filtering. For content filtering, a couple of use cases were
contributed. Service Providers may, from time to time, be requested
by law enforcement agencies to block access to particular sites such
as online betting and gambling, or access to dating sites. Content
Filtering may also happen at the endpoints or at the edge of
enterprise networks. This section is intended to merely document
this current practice by operators and the effects of encryption on
the practice.
Content filtering motivations vary and in the mobile network
usually occurs in the core network. A proxy is installed which
analyses the transport metadata of the content users are viewing and
either filters content based on a blacklist of sites or based on the
user's pre-defined profile (e.g. for age sensitive content).
Although filtering can be done by many methods one common method
occurs when a DNS lookup of a hostname in a URL which appears on a
government or recognized block-list (
aims to address this). The subsequent requests to that domain will
be re-routed to a proxy which checks whether the full URL matches a
blocked URL on the list, and will return a 404 if a match is found.
All other requests should complete.
See Section 7 for more information on "Encryption Impact on
Mobility Network Optimizations and New Services".
Another form of content filtering is called parental control,
where some users are deliberately denied access to age-sensitive
content as a feature to the service subscriber. Some sites involve a
mixture of universal and age-sensitive content and filtering
software. In these cases, more granular (application layer) metadata
may be used to analyze and block traffic. Methods that accessed
cleartext application-layer metadata no longer work when sessions
are encrypted. This type of granular filtering could occur at the
endpoint, however the ability to efficiently provide this as a
service without new efficient management solutions for end point
solutions impacts providers.
There are cases (beyond parental control) when a mobile network
service provider redirects customer requests for content:
The mobile network service provider is performing the
accounting and billing for the content provider, and the
customer has not (yet) purchased the requested content.
Further content may not be allowed as the customer has
reached their usage limit and needs to purchase additional data
service.
Currently, the mobile network service provider redirects
the customer using HTTP redirect to a page which educates the
customer on the reason for the blockage and provide steps to
proceed. Once the HTTP header and content are encrypted, the Mobile
carrier loses the option to intercept the traffic and perform an
HTTP redirect. With current solution options, this leaves only the
option to block the customer's request and cause a bad customer
experience until the blocking reason can be conveyed by some other
means. The customer may need to call customer care to find out the
reason, both an inconvenience to the customer and additional
overhead to the mobile network service provider.
Where network load balancers have been configured to route
according to application-layer semantics, an encrypted payload is
effectively invisible. This has resulted in practices of
intercepting TLS in front of load balancers to regain that
visibility, but at a cost to security and privacy.
Approved access to a network is a prerequisite to requests for
Internet traffic - hence network access, including any
authentication and authorization, is not impacted by encryption.
Cellular networks often sell tariffs that allow free-data access
to certain sites, known as 'zero rating'. A session to visit such a
site incurs no additional cost or data usage to the user. This
feature may be impacted if encryption hides the details of the
content domain from the network. This topic and related material are
described further in the Appendix.
Mobile networks (and usually ISPs) operate under the regulations
of their licensing government authority. These regulations include
Lawful Intercept, adherence to Codes of Practice on content
filtering, and application of court order filters.
These functions are impacted by encryption, typically by allowing
a less granular means of implementation. The enforcement of any Net
Neutrality regulations is unlikely to be affected by content being
encrypted. The IETF's Policy on Wiretapping can be found in , which does not support wiretapping in
standards.
The policy of some mobile network service providers to deploy
Application Layer Gateways (ALG). Section 2.9 of describes the role of ALG and their interaction
with NAT and/or the application payload. ALG are deployed to provide
connectivity across Network Address Translators (NAT), Firewalls,
and/or Load Balancers for specific applications the mobile network
providers choose to support. One example is a video application that
uses the Real Time Session Protocol (RTSP)
primary stream as a means to identify related Real Time
Protocol/Real Time Control Protocol (RTP/RTCP) flows at set-up. The ALG relies on the 5-tuple
flow information derived from RTSP to provision NAT or other middle
boxes and provide connectivity. Implementations vary, and two
examples follow:
Parse the content of the RTSP stream and identify the 5-tuple
of the supporting streams as they are being negotiated.
Intercept and modify the 5-tuple information of the
supporting media streams as they are being negotiated on the
RTSP stream, which is more intrusive to the media streams.
HTTP header insertion (see section 3.2.1 of ) has been a mechanism for the mobile carrier to
provide “allowed” (Non-Customer Proprietary Network
Information) subscriber information to third parties or other
internal systems . Third parties can in turn
provide customized service, or use it to bill the customer or
allow/block selective content. This 'header-enrichment' method is
also used within the mobile network service provider to pass
information internally between sub-systems, thus keeping the
internal systems loosely-coupled. With encryption, the mobile
network service provider loses the capability to include any
information in the header itself, but this is one motivation for
encryption.
Network operators are often the first ones called upon to
investigate any application problems (e.g., "my HD video is choppy").
By investigating packet loss (from sequence and acknowledgement
numbers), round-trip-time (from TCP timestamp options or
application-layer transactions, e.g., DNS or HTTP response time),
receive-window size, packet corruption (from checksum verification),
inefficient fragmentation, or application-layer problems, the operator
can narrow the problem to a portion of the network, server overload,
client or server misconfiguration, etc. Network operators may also be
able to identify the presence of attack traffic as not conforming to
the application the user claims to be using.
One way of quickly excluding the network as the bottleneck during
troubleshooting is to check whether the speed is limited by the
endpoints. For example, the connection speed might instead be limited
by suboptimal TCP options, the sender's congestion window, the sender
temporarily running out of data to send, the sender waiting for the
receiver to send another request, or the receiver closing the receive
window.
Packet captures and protocol-dissecting analyzers have been
important tools. Automated monitoring has also been used to
proactively identify poor network conditions, leading to maintenance
and network upgrades before user experience declines. For example,
findings of loss and jitter in VoIP traffic can be a predictor of
future customer dissatisfaction, or increases in DNS response time can
generally make interactive web browsing appear sluggish.
When utilizing increased encryption, application server operators
should expect to be called upon more frequently to diagnose problems,
and should consider what tools they can put in the hands of their
clients or network operators.
Similar to DPI, the performance of some services can be more
efficiently managed and repaired when information on user transactions
is available to the service provider. It may be possible to continue
such monitoring activities without clear text access to the
application layers of interest, but inaccuracy will increase and
efficiency of repair activities will decrease. For example, an
application protocol error or failure would be opaque to network
troubleshooters when transport encryption is applied, making root
cause location more difficult and therefore increasing the
time-to-repair. Repair time directly reduces the availability of the
service, and availability is a key metric for Service Level Agreements
and subscription rebates. Also, there may be more cases of user
communication failures when the additional encryption processes are
introduced, leading to more customer service contacts and (at the same
time) less information available to network operations repair
teams.
It is important to note that the push for encryption by application
providers has been motivated by the application of the described
techniques. Some application providers have noted degraded performance
and/or user experience when network-based optimization or enhancement
of their traffic has occurred, and such cases may result in additional
operator troubleshooting, as well.
With the use of WebSockets , many forms of
communications (from isochronous/real-time to bulk/elastic file
transfer) will take place over HTTP port 80 or port 443, so only the
messages and higher-layer data will make application differentiation
possible. If the monitoring systems sees only "HTTP port 443", it
cannot distinguish application streams that would benefit from
priority queueing from others that would not.
Hosted environments have had varied requirements in the past for
encryption, with many businesses choosing to use these services
primarily for data and applications that are not business or privacy
sensitive. A shift prior to the revelations on surveillance/passive
monitoring began where businesses were asking for hosted environments to
provide higher levels of security so that additional applications and
service could be hosted externally. Businesses understanding the threats
of monitoring in hosted environments only increased that pressure to
provide more secure access and session encryption to protect the
management of hosted environments as well as for the data and
applications.
Hosted environments may have multiple levels of management access,
where some may be strictly for the Hosting SP (infrastructure that may
be shared among customers) and some may be accessed by a specific
customer for application management. In some cases, there are multiple
levels of hosting service providers, further complicating the security
of management infrastructure and the associated requirements.
Hosting service provider management access is typically segregated
from other traffic with a control channel and may or may not be
encrypted depending upon the isolation characteristics of the
management session. Customer access may be through a dedicated
connection, but discussion for that connection method is
out-of-scope.
Application Service Providers may offer content-level monitoring
options to detect intellectual property leakage, or other attacks. The
use of session encryption will prevent Data Leakage Protection (DLP)
used on the session streams from accessing content to search on
keywords or phases to detect such leakage. DLP is often used to
prevent the leakage of Personally Identifiable Information (PII) as
well as financial account information, Personal Health Information
(PHI), and Payment Card Information (PCI). If session encryption is
terminated at a gateway prior to accessing these services, DLP on
session data can still be performed. The decision of where to
terminate encryption to hosted environments will be a risk decision
made between the application service provider and customer
organization according to their priorities. DLP can be performed at
the server for the hosted application and on an end users system in an
organization as alternate or additional monitoring points of content,
however this is not frequently done in a service provider
environment.
Application service providers, by their very nature, control the
application endpoint. As such, much of the information gleaned from
sessions are still available on that endpoint. Additionally, a gap may
exist in the logging and debugging capabilities of the applications
that led to the use of accessing data in transport for some of the
monitoring applications.
Overlay networks (e.g. VXLAN, Geneve, etc.) may be used to indicate
desired isolation, but this is not sufficient to prevent deliberate
attacks that are aware of the use of the overlay network. It is
possible to use an overlay header in combination with IPsec, but this
adds the requirement for authentication infrastructure and may reduce
packet transfer performance. Additional extension mechanisms to
provide integrity and/or privacy protections are being investigated
for overlay encapsulations. Section 7 of [RFC7348] describes some of
the security issues possible when deploying VXLAN on Layer 2 networks.
Rogue endpoints can join the multicast groups that carry broadcast
traffic, for example.
Hosted applications that allow some level of customer management
access may also require monitoring by the hosting service provider.
The monitoring needs could include access control restrictions such
as authentication, authorization, and accounting for filtering and
firewall rules to ensure they are continuously met. Customer access
may occur on multiple levels, including user-level and
administrative access. The hosting service provider may need to
monitor access either through session monitoring or log evaluation
to ensure security service level agreements (SLA) for access
management are met. The use of session encryption to access hosted
environments limits access restrictions to the metadata described
below. Monitoring and filtering may occur at an:
IP-level with source and destination IP
addresses alone, or
IP and protocol-level with source IP
address, destination IP address, protocol number, source port
number, and destination port number.
Session encryption at the application level, TLS for example,
currently allows access to the 5-tuple. IP-level encryption, such as
IPsec in tunnel mode prevents access to the original 5-tuple and may
limit the ability to restrict traffic via filtering techniques. This
shift may not impact all hosting service provider solutions as
alternate controls may be used to authenticate sessions or access
may require that clients access such services by first connecting to
the organization before accessing the hosted application. Shifts in
access may be required to maintain equivalent access control
management. Logs may also be used for monitoring that access control
restrictions are met, but would be limited to the data that could be
observed due to encryption at the point of log generation. Log
analysis is out of scope for this document.
The following observations apply to any IT organization that is
responsible for delivering services, whether to third-parties, for
example as a web based service, or to internal customers in an
enterprise, e.g. a data processing system that forms a part of the
enterprise’s business.
Organizations responsible for the operation of a data center have
many processes which access the contents of IP packets (passive
methods of measurement, as defined in ).
These processes are typically for service assurance or security
purposes as part of their data center operations.
Examples include:
- Network Performance Monitoring/Application Performance
Monitoring
- Intrusion defense/prevention systems
- Malware detection
- Fraud Monitoring
- Application DDOS protection
- Cyber-attack investigation
- Proof of regulatory compliance
Many application service providers simply terminate
sessions to/from the Internet at the edge of the data center in the
form of SSL/TLS offload in the load balancer. Not only does this
reduce the load on application servers, it simplifies the processes
to enable monitoring of the session content.
However, in some situations, encryption deeper in the data center
may be necessary to protect personal information or in order to meet
industry regulations, e.g. those set out by the Payment Card
Industry (PCI). In such situations, various methods have been used
to allow service assurance and security processes to access
unencrypted data. These include SSL/TLS decryption in dedicated
units, which then forward packets to 'trusted' tools, or by
real-time or post-capture decryption in the tools themselves. The
use of tools that perform SSL/TLS decryption are impacted by the
increased use of encryption that prevents interception. Alternate
methods to acheive the goals of these functions may be necessary and
in some cases, the functions may no longer persist in a pervasively
encrypted Internet.
Data center operators may also maintain packet recordings in
order to be able to investigate attacks, breach of internal
processes, etc. In some industries, organizations may be legally
required to maintain such information for compliance purposes.
Investigations of this nature have used access to the unencrypted
contents of the packet. Alternate methods to investigate attacks or
breach of process will rely on endpoint information, such as logs.
As noted previously, logs are often lacking in the information
provided and is seen as a current gap hence the problem for those
relying on session access.
Organizations are increasingly using hosted applications rather
than in house solutions that require maintenance of equipment and
software. Examples include Enterprise Resource Planning (ERP)
solutions, payroll service, time and attendance, travel and expense
reporting among others. Organizations may require some level of
management access to these hosted applications and will typically
require session encryption or a dedicated channel for this
activity.
In other cases, hosted applications may be fully managed by a
hosting service provider with service level agreement expectations for
availability and performance as well as for security functions
including malware detection. Due to the sensitive nature of these
hosted environments, the use of encryption is already prevalent. Any
impact may be similar to an enterprise with tools being used inside of
the hosted environment to monitor traffic. Additional concerns were
not reported in the call for contributions.
Performance, availability, and other aspects of a SLA are often
collected through passive monitoring. For example:
Availability: ability to establish connections with hosts to
access applications, and discern the difference between network
or host-related causes of unavailability.
Performance: ability to complete transactions within a target
response time, and discern the difference between network or
host-related causes of excess response time.
Here, as with all passive monitoring, the accuracy of inferences
are dependent on the cleartext information available, and encryption
would tend to reduce the information and therefore, the accuracy of
each inference. Passive measurement of some metrics will be
impossible with encryption that prevents inferring packet
correspondence across multiple observation points, such as for
packet loss metrics.
Until application logging is sufficient, the ability to make
accurate inferences in an environment with increased encryption will
remain a gap.
Mail (application) service providers vary in what services they
offer. Options may include a fully hosted solution where mail is
stored external to an organization's environment on mail service
provider equipment or the service offering may be limited to monitor
incoming mail to remove SPAM [Section 5.1], malware [Section 5.6],
and phishing attacks [Section 5.3] before mail is directed to the
organization's equipment. In both of these cases, content of the
messages and headers is monitored to detect SPAM, malware, phishing,
and other messages that may be considered an attack.
STARTTLS ought have zero effect on anti-SPAM efforts for SMTP
traffic. Anti-SPAM services could easily be performed on an SMTP
gateway, eliminating the need for TLS decryption services. The
impact to Anti-SPAM service providers should be limited to a change
in tools, where middle boxes were deployed to perform these
functions.
Many efforts are emerging to improve user-to-user encryption to
protect end user's privacy. PGP may be a front runner, and there are
other efforts ranging from proprietary to open source ones like
"Dark Mail".
Numerous service offerings exist that provide hosted storage
solutions. This section describes the various offerings and details
the monitoring for each type of service and how encryption may impact
the operational and security monitoring performed.
Trends in data storage encryption for hosted environments include a
range of options. The following list is intentionally high-level to
describe the types of encryption used in coordination with data
storage that may be hosted remotely, meaning the storage is physically
located in an external data center requiring transport over the
Internet. Options for monitoring will vary with both approaches from
what may be done today.
For higher security and/or privacy of data and applications,
options that provide end-to-end encryption of the data from the
users desktop or server to the storage platform may be preferred.
With this description, host level encryption includes any solution
that encrypts data at the object level, not transport level.
Encryption of data may be performed with libraries on the system or
at the application level, which includes file encryption services
via a file manager. Host-level encryption is useful when data
storage is hosted, or scenarios when storage location is determined
based on capacity or based on a set of parameters to automate
decisions. This could mean that large data sets accessed
infrequently could be sent to an off-site storage platform at an
external hosting service, data accessed frequently may be stored
locally, or the decision could be based on the transaction type.
Host-level encryption is grouped separately for the purpose of this
document as the monitoring needs as this data can be stored in
multiple locations including off-site remote storage platforms. If
session encryption is used, the protocol is likely to be TLS.
The general monitoring needs of hosted storage solutions that
use host-level (object) encryption is described in this
subsection. Solutions might include backup services and external
storage services, such as those that burst data that exceeds
internal limits on occasion to external storage platforms operated
by a third party.
Monitoring of data flows to hosted storage solutions is
performed for security and operational purposes. The security
monitoring may be to detect anomalies in the data flows that could
include changes to destination, the amount of data transferred, or
alterations in the size and frequency of flows. Operational
considerations include capacity and availability monitoring.
There are multiple ways to achieve full disk encryption for
stored data. Encryption may be performed on data to be stored while
in transit close to the storage media with solutions like Controller
Based Encryption (CBE) or in the drive system with Self-Encrypting
Drives (SED). Session encryption is typically coupled with
encryption of these data at rest (DAR) solutions to also protect
data in transit. Transport encryption is likely via TLS.
The general monitoring needs for transport of data to storage
platforms, where object level encryption is performed close to or
on the storage platform are similar to those described in the
section on Monitoring for Hosted Storage. The primary difference
for these solutions is the possible exposure of sensitive
information, which could include privacy related data, financial
information, or intellectual property if session encryption via
TLS is not deployed. Session encryption is typically used with
these solutions, but that decision would be based on a risk
assessment. There are use cases where DAR or disk-level encryption
is required. Examples include preventing exposure of data if
physical disks are stolen or lost.
Storage services also include data replication which may occur
between data centers and may leverage Internet connections to tunnel
traffic. The traffic may use iSCSI or FC/IP
encapsulated in IPsec. Either transport or
tunnel mode may be used for IPsec depending upon the termination
points of the IPsec session, if it is from the storage platform
itself or from a gateway device at the edge of the data center
respectively.
The general monitoring needs for data replication are described
in this subsection.
Monitoring of data flows between data centers may be performed
for security and operational purposes and would typically
concentrate more on the operational needs since these flows are
essentially virtual private networks (VPN) between data centers.
Operational considerations include capacity and availability
monitoring. The security monitoring may be to detect anomalies in
the data flows, similar to what was described in the "Monitoring
for Hosted Storage Section".
Encryption of network traffic within the private enterprise is a
growing trend, particularly in industries with audit and regulatory
requirements. Some enterprise internal networks are almost completely
TLS and/or IPsec encrypted.
For each type of monitoring, different techniques and access to parts
of the data stream are part of current practice. As we transition to an
increased use of encryption, alternate methods of monitoring for
operational purposes may be necessary to reduce the practice of breaking
encryption and thus privacy of users (other policies may apply in some
enterprise settings).
Large corporate enterprises are the owners of the platforms, data,
and network infrastructure that provide critical business services to
their user communities. As such, these enterprises are responsible for
all aspects of the performance, availability, security, and quality of
experience for all user sessions. These responsibilities break down
into three basic areas:
Security Monitoring and Control
Application Performance Monitoring and Reporting
Network Diagnostics and Troubleshooting
In each of the above areas, technical support teams utilize
collection, monitoring, and diagnostic systems. Some organizations
currently use attack methods such as replicated TLS server RSA private
keys to decrypt passively monitored copies of encrypted TLS packet
streams.
For an enterprise to avoid costly application down time and deliver
expected levels of performance, protection, and availability, some
forms of traffic analysis sometimes including examination of packet
payloads are currently used.
Enterprise users are subject to the policies of their
organization and the jurisdictions in which the enterprise operates.
As such, proxies may be in use to:
intercept outbound session traffic to monitor for
intellectual property leakage (by users or more likely these
days through malware and trojans),
detect viruses/malware entering the network via email or web
traffic,
detect malware/Trojans in action, possibly connecting to
remote hosts,
detect attacks (Cross site scripting and other common web
related attacks),
track misuse and abuse by employees,
restrict the types of protocols permitted to/from the entire
corporate environment,
detect and defend against Internet DDoS attacks, including
both volumetric and layer 7 attacks.
A significant portion of malware hides its activity within
TLS or other encrypted protocols. This includes lateral movement,
Command and Control, and Data Exfiltration. Detecting these
functions are important to effective monitoring and mitigation of
malicious traffic, not limited to malware.
Security monitoring in the enterprise may also be performed at
the endpoint with numerous current solutions that mitigate the same
problems as some of the above mentioned solutions. Since the
software agents operate on the device, they are able to monitor
traffic before it is encrypted, monitor for behavior changes, and
lock down devices to use only the expected set of applications.
Session encryption does not affect these solutions. Some might argue
that scaling is an issue in the enterprise, but some large
enterprises have used these tools effectively.
There are two main goals of monitoring:
Assess traffic volume on a per-application basis, for
billing, capacity planning, optimization of geographical
location for servers or proxies, and other goals.
Assess performance in terms of application response time and
user perceived response time.
Network-based Application Performance Monitoring tracks
application response time by user and by URL, which is the
information that the application owners and the lines of business
request. Content Delivery Networks (CDNs) add complexity in
determining the ultimate endpoint destination. By their very nature,
such information is obscured by CDNs and encrypted protocols --
adding a new challenge for troubleshooting network and application
problems. URL identification allows the application support team to
do granular, code level troubleshooting at multiple tiers of an
application.
New methodologies to monitor user perceived response time and to
separate network from server time are evolving. For example, the
IPv6 Destination Option Header (DOH) implementation of Performance
and Diagnostic Metrics (PDM) will provide this . Using PDM with IPSec
Encapsulating Security Payload (ESP) Transport Mode requires
placement of the PDM DOH within the ESP encrypted payload to avoid
leaking timing and sequence number information that could be useful
to an attacker. Use of PDM DOH also may introduce some security
weaknesses, including a timing attack, as described in Section 7 of
. For these and other
reasons, requires
that the PDM DOH option be explicitly turned on by administrative
action in each host where this measurement feature will be used.
One primary key to network troubleshooting is the ability to
follow a transaction through the various tiers of an application in
order to isolate the fault domain. A variety of factors relating to
the structure of the modern data center and the modern multi-tiered
application have made it difficult to follow a transaction in
network traces without the ability to examine some of the packet
payload. Alternate methods, such as log analysis need improvement to
fill this gap.
Content Delivery Networks (CDNs) and NATs and Network Address
and Port Translators (NAPT) obscure the ultimate endpoint
designation (See for types of address
sharing and a list of issues). Troubleshooting a problem for a
specific end user requires finding information such as the IP
address and other identifying information so that their problem
can be resolved in a timely manner.
NAT is also frequently used by lower layers of the data center
infrastructure. Firewalls, Load Balancers, Web Servers, App
Servers, and Middleware servers all regularly NAT the source IP of
packets. Combine this with the fact that users are often allocated
randomly by load balancers to all these devices, the network
troubleshooter is often left with very few options in today's
environment due to poor logging implementations in applications.
As such, network troubleshooting is used to trace packets at a
particular layer, decrypt them, and look at the payload to find a
user session.
This kind of bulk packet capture and bulk decryption is
frequently used when troubleshooting a large and complex
application. Endpoints typically don't have the capacity to handle
this level of network packet capture, so out-of-band networks of
robust packet brokers and network sniffers that use techniques
such as copies of TLS RSA private keys accomplish this task
today.
There is an experimental TCP option for host identification
which can mitigate the complications with address sharing devices
(including CDN border nodes) by supplying the original IP address
.
TCP Pipelining/Session Multiplexing used mainly by middle boxes
today allow for multiple end user sessions to share the same TCP
connection. Today's network troubleshooter often relies upon
session decryption to tell which packet belongs to which end user
as the logs are currently inadequate for the analysis
performed.
With the advent of HTTP2, session multiplexing will be used
ubiquitously, both on the Internet and in the private data
center.
When an application server makes an HTTP service call to back
end services on behalf of a user session, it uses a completely
different URL and a completely different TCP connection.
Troubleshooting via network trace involves matching up the user
request with the HTTP service call. Some organizations do this
today by decrypting the TLS packet and inspecting the payload.
Logging has not been adequate for their purposes.
Many applications use text formats such as XML to transport
data or application level information. When transaction failures
occur and the logs are inadequate to determine the cause, network
and application teams work together, each having a different view
of the transaction failure. Using this troubleshooting method, the
network packet is correlated with the actual problem experienced
by an application to find a root cause. The inability to access
the payload prevents this method of troubleshooting.
Corporate networks commonly monitor outbound session traffic to
detect or prevent attacks as well as to guarantee service level
expectations. In some cases, alternate options are available when
encryption is in use, but techniques like that of data leakage
prevention tools at a proxy would not be possible if encrypted traffic
can not be intercepted, encouraging alternate options such as
performing these functions at the edge.
DLP tools intercept traffic at the Internet gateway or proxy
services with the ability to man-in-the-middle (MiTM) encrypted
session traffic (HTTP/TLS). These tools may use key words important to
the enterprise including business sensitive information such as trade
secrets, financial data, personally identifiable information (PII), or
personal health information (PHI). Various techniques are used to
intercept HTTP/TLS sessions for DLP and other purposes, and are
described in "Summarizing Known Attacks on TLS and DTLS" . Note: many corporate policies allow access to
personal financial and other sites for users without interception.
Monitoring traffic patterns for anomalous behavior such as
increased flows of traffic that could be bursty at odd times or flows
to unusual destinations (small or large amounts of traffic) is common.
This traffic may or may not be encrypted and various methods of
encryption or just obfuscation may be used.
Restrictions on traffic to approved sites: Web proxies are
sometimes used to filter traffic, allowing only access to well-known
sites found to be legitimate and free of malware on last check by a
proxy service company. This type of restriction is usually not
noticeable in a corporate setting, but may be to those in research who
are unable to access colleague's individual sites or new web sites
that have not yet been screened. In situations where new sites are
required for access, they can typically be added after notification by
the user or proxy log alerts and review. Home mail account access may
be blocked in corporate settings to prevent another vector for malware
to enter as well as for intellectual property to leak out of the
network. This method remains functional with increased use of
encryption and may be more effective at preventing malware from
entering the network. Web proxy solutions monitor and potentially
restrict access based on the destination URL or the DNS name. A
complete URL may be used in cases where access restrictions vary for
content on a particular site or for the sites hosted on a particular
server.
Desktop DLP tools are used in some corporate environments as well.
Since these tools reside on the desktop, they can intercept traffic
before it is encrypted and may provide a continued method of
monitoring intellectual property leakage from the desktop to the
Internet or attached devices.
DLP tools can also be deployed by Network Service providers, as
they have the vantage point of monitoring all traffic paired with
destinations off the enterprise network. This makes an effective
solution for enterprises that allow "bring-your-own" devices when the
traffic is not encrypted and devices that do not fit the desktop
category, but are used on corporate networks nonetheless.
Enterprises may wish to reduce the traffic on their Internet access
facilities by monitoring requests for within-policy content and
caching it. In this case, repeated requests for Internet content
spawned by URLs in e-mail trade newsletters or other sources can be
served within the enterprise network. Gradual deployment of end to end
encryption would tend to reduce the cacheable content over time, owing
to concealment of critical headers and payloads. Many forms of
enterprise performance management and optimization based on monitoring
(DPI) would suffer the same fate.
Effective incident response today requires collaboration at Internet
scale. This section will only focus on efforts of collaboration at
Internet scale that are dedicated to specific attack types. They may
require new monitoring and detection techniques in an increasingly
encrypted Internet. As mentioned previously, some service providers have
been interfering with STARTTLS to prevent session encryption to be able
to perform functions they are used to (injecting ads, monitoring, etc.).
By detailing the current monitoring methods used for attack detection
and response, this information can be used to devise new monitoring
methods that will be effective in the changed Internet via collaboration
and innovation.
The largest operational effort to prevent mail abuse is through the
Messaging, Malware, Mobile Anti-Abuse Working Group (M3AAWG). Mail abuse is combated directly with mail
administrators who can shut down or stop continued mail abuse
originating from large scale providers that participate in using the
Abuse Reporting Format (ARF) agents standardized in the IETF , , , , , , and . The ARF agent directly reports abuse messages to
the appropriate service provider who can take action to stop or
mitigate the abuse. Since this technique uses the actual message, the
use of SMTP over TLS between mail gateways will not effect its
usefulness. As mentioned previously, SMTP over TLS only protects data
while in transit and the messages may be exposed on mail servers or
mail gateways if a user-to-user encryption method is not used. Current
user-to-user message encryption methods on email (S/MIME and PGP) do
not encrypt the email header information used by ARF and the service
provider operators in their abuse mitigation efforts.
Response to Denial of Service (DoS) attacks are typically
coordinated by the SP community with a few key vendors who have tools
to assist in the mitigation efforts. Traffic patterns are determined
from each DoS attack to stop or rate limit the traffic flows with
patterns unique to that DoS attack.
Data types used in monitoring traffic for DDoS are described in the
DDoS Open Threat Signaling (DOTS) working group documents in
development.
Data types used in DDoS attacks have been detailed in the IODEF
Guidance draft , Appendix
A.2, with the help of several members of the service provider
community. The examples provided are intended to help identify the
useful data in detecting and mitigating these attacks independent of
the transport and protocol descriptions in the drafts.
Investigations and response to phishing attacks follow well-known
patterns, requiring access to specific fields in email headers as well
as content from the body of the message. When reporting phishing
attacks, the recipient has access to each field as well as the body to
make content reporting possible, even when end-to-end encryption is
used. The email header information is useful to identify the mail
servers and accounts used to generate or relay the attack messages in
order to take the appropriate actions. The content of the message
often contains an embedded attack that may be in an infected file or
may be a link that results in the download of malware to the users
system.
Administrators often find it helpful to use header information to
track down similar message in their mail queue or users inboxes to
prevent further infection. Combinations of To:, From:, Subject:,
Received: from header information might be used for this purpose.
Administrators may also search for document attachments of the same
name, size, or containing a file with a matching hash to a known
phishing attack. Administrators might also add URLs contained in
messages to block lists locally or this may also be done by browser
vendors through larger scales efforts like that of the Anti-Phishing
Working Group (APWG).
A full list of the fields used in phishing attack incident response
can be found in RFC5901. Future plans to increase privacy protections
may limit some of these capabilities if some email header fields are
encrypted, such as To:, From:, and Subject: header fields. This does
not mean that those fields should not be encrypted, only that we
should be aware of how they are currently used.
Some products protect users from phishing by maintaining lists of
known phishing domains (such as misspelled bank names) and blocking
access. This can be done by observing DNS, clear-text HTTP, or SNI in
TLS, in addition to analyzing email. Alternate options to detect and
prevent phishing attacks may be needed. More recent examples of data
exchanged in spear phishing attacks has been detailed in the IODEF
Guidance draft , Appendix
A.3.
Botnet detection and mitigation is complex and may involve hundreds
or thousands of hosts with numerous Command and Control (C&C)
servers. The techniques and data used to monitor and detect each may
vary. Connections to C&C servers are typically encrypted,
therefore a move to an increasingly encrypted Internet may not affect
the detection and sharing methods used.
Malware monitoring and detection techniques vary. As mentioned in
the enterprise section, malware monitoring may occur at gateways to
the organization analyzing email and web traffic. These services can
also be provided by service providers, changing the scale and location
of this type of monitoring. Additionally, incident responders may
identify attributes unique to types of malware to help track down
instances by their communication patterns on the Internet or by
alterations to hosts and servers.
Data types used in malware investigations have been summarized in
an example of the IODEF Guidance draft , Appendix A.1.
The IETF has reacted to spoofed source IP address-based attacks,
recommending the use of network ingress filtering and the unicast Reverse Path Forwarding (uRPF)
mechanism . But uRPF suffers from limitations
regarding its granularity: a malicious node can still use a spoofed IP
address included inside the prefix assigned to its link. The Source
Address Validation Improvements (SAVI) mechanisms try to solve this
issue. Basically, a SAVI mechanism is based on the monitoring of a
specific address assignment/management protocol (e.g., SLAAC , SEND , DHCPv4/v6 ) and, according to this
monitoring, set-up a filtering policy allowing only the IP flows with
a correct source IP address (i.e., any packet with a source IP
address, from a node not owning it, is dropped). The encryption of
parts of the address assignment/management protocols, critical for
SAVI mechanisms, can result in a dysfunction of the SAVI
mechanisms.
Although incident response work will continue, new methods to
prevent system compromise through security automation and continuous
monitoring may provide alternate approaches
where system security is maintained as a preventative measure.
This section describes specific techniques used in monitoring
applications that may apply to various network types.
When initiating the TLS handshake, the Client may provide an
extension field (server_name) which indicates the server to which it
is attempting a secure connection. TLS SNI was standardized in 2003 to
enable servers to present the "correct TLS certificate" to clients in
a deployment of multiple virtual servers hosted by the same server
infrastructure and IP-address. Although this is an optional extension,
it is today supported by all modern browsers, web servers and
developer libraries. Akamai reports that
many of their customer see client TLS SNI usage over 99%.
It should be noted that HTTP/2 introduces the
Alt-SVC method for upgrading the connection from HTTP/1 to either
unencrypted or encrypted HTTP/2. If the initial HTTP/1 request is
unencrypted, the destination alternate service name can be identified
before the communication is potentially upgraded to encrypted HTTP/2
transport. HTTP/2 requires the TLS implementation to support the
Server Name Indication (SNI) extension (see section 9.2 of ).
This information is only visible if the client is populating the
Server Name Indication extension. This need not be done, but may be
done as per TLS standard and as stated above this has been implemented
by all major browsers. Therefore, even if existing network filters
look out for seeing a Server Name Indication extension, they may not
find one. The per-domain nature of SNI may not reveal the specific
service or media type being accessed, especially where the domain is
of a provider offering a range of email, video, Web pages etc. For
example, certain blog or social network feeds may be deemed 'adult
content', but the Server Name Indication will only indicate the server
domain rather than a URL path.
ALPN is a TLS extension which may be used to indicate the
application protocol within the TLS session. This is likely to be of
more value to the network where it indicates a protocol dedicated to a
particular traffic type (such as video streaming) rather than a
multi-use protocol. ALPN is used as part of HTTP/2 'h2', but will not
indicate the traffic types which may make up streams within an HTTP/2
multiplex.
The content length of encrypted traffic is effectively the same as
the cleartext. Although block ciphers utilise padding this makes a
negligible difference. Bitrate and pacing are generally application
specific, and do not change much when the content is encrypted.
Multiplexed formats (such as HTTP/2 and QUIC) may however incorporate
several application streams over one connection, which makes the
bitrate/pacing no longer application-specific.
This Appendix considers the effects of transport level encryption on
existing forms of mobile network optimization techniques, as well as
potential new services. The material in this Appendix assumes
familiarity with mobile network concepts, specifications, and
architectures. Readers who need additional background should start with
the 3GPP's web pages on various topics of interest, especially the article on LTE. 3GPP provides a
mapping between their expanding technologies and the different series of
technical specifications . 3GPP also has a
canonical specification of their vocabulary, definitions, and acronyms
, as does the RFC Editor for abbreviations .
The stream of TCP ACKs that flow from a receiver of a byte stream
using TCP for reliability, flow-control, and NAT/firewall transversal
is called an ACK stream. The ACKs contain segment numbers that confirm
successful transmission and their RTT, or indicate packet loss
(duplicate ACKs). If this view of progress of stream transfer is lost,
then the mobile network has greatly reduced ability to monitor
transport layer performance. When the ACK stream is encrypted, it
prevents the following mobile network features from operating:
Measurement of Network Segment (Sector, eNodeB (eNB) etc.)
characterization KPIs (Retransmissions, packet drops, Sector
Utilization Level etc.), estimation of User/Service KQIs at
network edges for circuit emulation (CEM), and mitigation methods.
The active services per user and per sector are not visible to a
server that only services Internet Access Point Names (APN), and
thus could not perform mitigation functions based on network
segment view.
Retransmissions by performance-enhancing proxies (see section
2.1.1 of and section 3.5 of
)at network edges that
improve live transmission over long delay, capacity-varying
networks.
Content replication near the network edge (for example live
video, DRM protected content) to maximize QOE. Replicating every
stream through the transit network increases backhaul cost for
live TV. There are alternate approaches such as blind caches being explored to allow caching of
encrypted content.
Ability to deploy trusted proxies that reduce control
round-trip time (RTT) between the TCP transmitter and receiver.
The RTT determines how quickly a user’s attempt to cancel a
video is recognized (how quickly the traffic is stopped, thus
keeping un-wanted video packets from entering the radio scheduler
queue).
Performance-enhancing proxy with low RTT determines the
responsiveness of TCP flow control, and enables faster adaptation
in a delay & capacity varying network due to user mobility.
Low RTT permits use of a smaller send window, which makes the flow
control loop more responsive to changing mobile network
conditions.
When the Transport Header is encrypted, it prevents the following
mobile network features from operating:
Application-type-aware network edge (middlebox) that could
control pacing, limit simultaneous HD videos, prioritize active
videos against new videos, etc.
For Self Organizing Networks (3GPP SON) – intelligent SON
workflows such as content-ware MLB (Mobility Load Balancing)
For User Plane Congestion Management (3GPP UPCON) –
ability to understand content and manage network during
congestion. Mitigating techniques such as deferred download,
off-peak acceleration, and outbound roamers.
Reduces the benefits IP/DSCP-based transit network delivery
optimizations; since the multiple applications are multiplexed
within the same 5-tuple transport connection; a reasonable
assumption is that the DSCP markings would be withheld from the
outer IP header to further obscure which packets belong to each
application flow.
Advance notification for dense data usages – If the
application types are visible, transit network element could warn
(ahead of usage) that the requested service consumes user plan
limits, and transmission could be terminated. Without such
visibility the network might have to continue the operation and
stop the operation after the limit, because partially loaded
content wastes resources and may not be usable by the client thus
increasing customer complaints. Content publisher will not know
user-service plans, and Network Edge would not know data transfer
lengths before large object is requested.
This section describes some new/emerging mobile services and how
they might be affected with transport encryption:
Content/Application based Prioritization of Over-the-Top (OTT)
services – each application-type or service has different
delay/loss/throughput expectations, and each type of stream will
be unknown to an edge device if encrypted; this impedes
dynamic-QoS adaptation.
Rich Communication Services (3GPP-RCS) using different Quality
Class Indicators (QCIs in LTE) – Operators offer different
QoS classes for value-added services. The QCI type is visible in
RAN control plane and invisible in user plane, thus the QCI cannot
be set properly when the application -type is unknown.
Enhanced Multimedia Broadcast/Multicast Services (3GPP eMBMS)
– trusted edge proxies facilitate delivering same stream to
different users, using either unicast or multicast depending on
channel conditions to the user.
The transport header encryption prevents trusted transit proxies.
It may be that the benefits of such proxies could be achieved by end
to end client & server optimizations and distribution using CDNs,
plus the ability to continue connections across different access
technologies (across dynamic user IP addresses). The following aspects
need to be considered in this approach:
In a wireless mobile network, the delay and channel capacity
per user and sector varies due to coverage, contention, user
mobility, and scheduling balances fairness, capacity and service
QoE. If most users are at the cell edge, the controller cannot use
more complex QAM, thus reducing total cell capacity; similarly if
a UMTS edge is serving some number of CS-Voice Calls, the
remaining capacity for packet services is reduced.
Roamers: Mobile wireless networks service in-bound roamers
(Users of Operator A in a foreign operator Network B) by
backhauling their traffic though Operator B’s network to
Operator A’s Network and then serving through the P-Gateway
(PGW), General GPRS Support Node (GGSN), Content Distribution
Network (CDN) etc., of Operator A (User’s Home Operator).
Increasing window sizes to compensate for the path RTT will have
the limitations outlined earlier for TCP. The outbound roamer
scenario has a similar TCP performance impact.
Issues in deploying CDNs in RAN: Decreasing Client-Server
control loop requires deploying CDNs/Cloud functions that
terminate encryption closer to the edge. In Cellular RAN, the user
IP traffic is encapsulated into GPSR Tunneling Protocol-User Plane
(GTP-U in UMTS and LTE) tunnels to handle user mobility; the
tunnels terminate in APN/GGSN/PGW that are in central locations.
One user's traffic may flow through one or more APN’s (for
example Internet APN, Roaming APN for Operator X, Video-Service
APN, OnDeckAPN etc.). The scope of operator private IP addresses
may be limited to specific APN. Since CDNs generally operate on
user IP flows, deploying them would require enhancing them with
tunnel translation, etc., tunnel management functions.
While CDNs that de-encrypt flows or split-connection proxy
(similar to split-tcp) could be deployed closer to the edges to
reduce control loop RTT, with transport header encryption, such
CDNs perform optimization functions only for partner client flows;
thus content from Small-Medium Businesses (SMBs) would not get
such CDN benefits.
In the best case scenario, engineers and other innovators would work
to solve the problems at hand in new ways rather than prevent the use of
encryption. It will take time to devise alternate approaches to achieve
similar goals.
There has already been documented cases of service providers
preventing STARTTLS to prevent session
encryption negotiation on some session to inject a super cookie.
It is well known that national surveillance programs monitor traffic
for terrorism as Internet security practitioners
monitor for criminal activities. Governments vary on their balance
between monitoring versus the protection of user privacy, data, and
assets. Those that favor unencrypted access to data ignore the real need
to protect users identity, financial transactions and intellectual
property, which requires security and encryption to prevent crime. A
clear understanding of technology, encryption, and monitoring goals will
aid in the development of solutions to appropriately balance these with
privacy. As this understanding increases, hopefully the discussions will
improve and this draft is meant to help further the discussion.
Terrorists and criminals have been using encryption for many years.
The current push to increase encryption is aimed at increasing users
privacy. There is already protection in place for purchases, financial
transactions, systems management infrastructure, and intellectual
property although this too can be improved. The Opportunistic Security
(OS) efforts aim to increase the costs of
monitoring through the use of encryption that can be subject to active
attacks, but make passive monitoring broadly cost prohibitive. This is
meant to restrict monitoring to sessions where there is reason to have
suspicion.
Open questions: As the use of encryption increases, does passive
monitoring become limited to metadata analysis? What metadata should be
left in protocols as they evolve to also protect users privacy? Can we
make changes to protocols and message exchanges to alter the current
monitoring practices at least for operations and security
practitioners?
There are no additional security considerations as this is a summary
and does not include a new protocol or functionality.
This memo makes no requests of IANA.
Thanks to our reviewers, Natasha Rooney, Kevin Smith, Ashutosh Dutta,
Brandon Williams, Jean-Michel Combes, Nalini Elkins, Paul Barrett, Badri
Subramanyan, Igor Lubashev, Suresh Krishnan, Dave Dolson, Mohamed
Boucadair, and Stephen Farrell for their editorial and content
suggestions. Surya K. Kovvali provided material for the Appendix.
Electronic Frontier Foundation https://www.eff.org/
CAIDA [http://www.caida.org/data/overview/]
Communications Assistance for Law Enforcement Act
(CALEA)
Pub. L. No. 103-414, 108 Stat. 4279, codified at 47
USC 1001-1010
Telecommunications security; Lawful Interception (LI);
Requirements of Law Enforcement Agencies
ETSI TS 101 331 V1.1.1 (2001-08)
Messaging, Malware, Mobile Anti-Abuse Working Group (M3AAWG)
https://www.maawg.org/
ISPs Removing their Customers EMail Encryption
https://www.eff.org/deeplinks/2014/11/starttls-downgrade-attacks/
Acord BoF IETF95
https://www.ietf.org/proceedings/95/accord.html
10 Standards for Oversight and Transparency of National
Intelligence Services http://jnslp.com/
Surveillance, Vol. 8 No. 3
EFF Report on STARTTLS Downgrade Attacks
https://www.eff.org/deeplinks/2014/11/starttls-downgrade-attacks
3GPP Web pages on specific topics of interest
http://www.3gpp.org/technologies/95-keywords-acronyms
3GPP TR 21.905 V13.1.0 (2016-06) Vocabulary for 3GPP
Specifications
https://portal.3gpp.org/desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=558
RFC Editor Abbreviation List
https://www.rfc-editor.org/materials/abbrev.expansion.txt
Mapping between technologies and specifications
http://www.3gpp.org/technologies
Investigating Transparent Web Proxies in Cellular Networks,
Passive and Active Measurement Conference (PAM)
USC
Header Enrichment or ISP Enrichment? Emerging Privacy Threats
in Mobile Networks, Hot Middlebox’15, August 17-21 2015,
London, United Kingdom
ICSI Berkley
Security Automation and Continuous Monitoring (sacm) IETF
Working Group
DDoS Open Threat Signaling IETF
Working Group
Erik Nygren, personal reference